A paradoxical condition of addiction is that people persist in self-destructive behavior despite knowledge of the negative consequences. To better understand this paradox, dual process models of addiction have emerged that emphasize two interrelated systems: an impulsive or incentive salience system and a cognitive control system. The incentive salience system uses fast and automatic associations to evaluate stimuli for its emotional and motivational significance. The cognitive control system underlies slower, reflective processes to guide behavior, often overriding automatic action tendencies. As imbalances between the two systems increase, drug use becomes more involuntary and compulsive, leading to addiction. Cognitive approaches have shown that attention and approach biases to drugs and drug-related cues emerge after repeated drug use. There have been few attempts, however, to map the neurobiological basis of these attention and approach biases. Additionally, researchers have begun to implement bias retraining interventions in clinical settings, with initial results indicating such retraining procedures provide incremental improvement over traditional treatments that focus on cognitive control processes (e.g., CBT). Recently, researchers have begun to examine the validity of computer administered bias retraining treatments for SUDs, with initial positive results. We prose to examine the incremental benefit of bias retraining for SUDs while also examining the neurobiological mechanisms of change in these attention and approach. Specifically, to what extent can treatment affect changes in incentive salience and cognitive control processes associated with addiction? As these automatic processes occur on the order of fractions of a second, we will use sophisticated EEG/ERP methods (time-frequency, phase synchrony, source localization), as they provide the necessary temporal resolution and well-validated measures of electrocortical activity associated with the incentive salience and cognitive control systems. We also frame our analysis in the developmental context of adolescence. Adolescence is a period of substantial changes in incentive salience and cognitive control processes, while also the time when people tend to initiate substance use and early onset substance use disorders (SUDs) first emerge. Therefore, adolescence may be the key period to both study etiological mechanisms, and intervene to avert chronic SUDs. We will examine cognitive control and incentive salience functioning in adolescent patients in treatment for SUDs using EEG/ERP measures elicited in drug-specific and standard task paradigms. Patients will be randomly assigned to treatment as usual (TAU; focus on cognitive control) or TAU+ bias retraining (additional focus on attention and approach biases). Our specific aims are to (1) assess the utility of baseline EEG/ERP measures to predict severity and treatment response, (2) the efficacy of TAU+ bias retraining relative to TAU, and (3) whether EEG/ERP measures exhibit change as a function of treatment condition. We predict both treatment conditions will be associated with change in cognitive control, while TAU+ bias retraining will also be associated with change in incentive salience.